# Calibrating a capacitive moisture sensor

I am doing a project which require to build a moisture sensor to measure wood chip moisture content. I built the the cylinder and put all the capacitor plate around it and I was able to measure the capacitance of the wood chip. However, I do not know how to convert capacitance that i got to the moisture content. I did a lot of research online but they do not talk briefly about it. Is there any specific equation or way to convert capacitance to moisture content?

• It can (and should) be done empirically. You measure the moisture and the capacitance by some reliable tool. Repeating it for many different values, drawing a graph and deriving an approximation function. Excel can do it for you. Commented Mar 7, 2018 at 17:57
• You need a 2D graph such as Excel. Rows can be know capacitive values. Columns can be temperature ranges. Write in the best-guess reading for as much resolution as your willing to write.
– user105652
Commented Mar 15, 2018 at 0:56

There's not going to be an equation online for you to just plug in your values, as it's going to depend on your specific device, how you're measuring the capacitance, and probably even the wood you're using. Instead, you're going to need to calibrate it yourself, so you'd collect a bunch of points at known moisture values (ex. 0%, 10%, 20%, ...) and measure the capacitance for each. Assuming they form a smooth curve, you can then interpolate between capacitance/moisture values to get your reading

1. Prepare samples of wood-chip with, if possible, a range of moisture values that cover the range you expect to see in practice. This may involve drying some and steaming some or just leaving them outside, etc. Bag and label each sample.
2. Take a moisture content measurement of each sample using your sensor. Record the reading and the batch used.
3. Immediately weigh the sample. A digital kitchen scales should suffice. Record the readings, ($m_1$).
4. Dry the samples in the oven at a gentle heat until you think the are completely dry.
5. Weigh each sample again. Record the readings, ($m_2$).
6. Plot the results on graph paper. If you are a good lab technician your data points will align closely with the best-fit curve between the points.

Moisture content can be calculated as $\frac {m_1 - m_2}{m_2}\cdot 100$ (%).

Figure 1. Microwave Oven Drying, by Richard Jones makes it look easy.

Moisture Content by the Oven Dry Method may be of interest.

The best way to do this would be to use some sort of reference meter and several objects of known moisture levels in 5 or 10% steps and plot the capacitance read on each of them. To get the best accuracy, you'll need to do the tests at a stable ambient temperature then repeat the tests at different ambients. You will probably find a variation in your readings which you can also plot on your graph.

If you wanted to take it further, you could then try different materials and repeat all the tests again. Plugging your values into excel will give you (hopefully) a nice visual representation of your results which you can then use as a means to calibrate your device. Be sure to double check the results, don't just do it once. You need to make sure your results are repeatable if you want to use the data for calibration!

You need to buy/rent an appropriate sensor to calibrate your own. You soak your wood chips, then measure the moisture content using your calibrated sensor and the capacitance with your device. That gives you the first point. By example 1.000nF => 95% moisture. Then you dry your wood chip and measure with the sensor and your sensor. You get second point, by example 0.990nF => 5% moisture. You plot the line and you get moisture level as a function of capacitance.

In fact, it will be much more complicated: you may have non-linearity in your device forcing you tu use either a non-linear model or to take a lot of measurements (maybe 1% moisture apart). The temperature may offset your curve and so on. This is yours to try...